AKD ITS SELF-COKSEKVATIO^. 11 



explicitly directed to the parts of an object, those parts 

 become, in turn, mutually exclusive objects of attention. 



Now, as pure void, space can have no limits. For any 

 possible boundary of space could only be the limit between 

 the given space and another space on the other side of 

 the boundary. Any possible limited space must have 

 geometrical form. But every geometrical form is neces- 

 sarily bounded by surfaces. Nay, a surface is ever to be 

 regarded as a boundary in a two-fold sense, if we are to 

 accept the guidance of mathematicians by whom in gen- 

 eral, and by Professor Clifford in particular, a surface is 

 denned as ' ' the boundary between two adjacent portions 

 of space. "* 



But a real boundary that is, a surface constituting a 

 transition between two volumes distinguishable in quality 

 can have no reality for space as such, since space, merely 

 as space, possesses and can therefore present no positive 

 difference in quality by which one space or portion of 

 space can be distinguished from another. 



It is evident, therefore, that any supposed limit of space 

 could only be a limit in space, the limit having objective 

 reality only through the existence of some object occupy- 

 ing space. So that all talk of a possible "curvature of 

 space" is at once chargeable with confounding extension, 

 as the universal and purely negative possibility of all 

 physical modes of existence, with a particular, positive, 

 material, extended object that might (and must) exist in 

 space, but could never coalesce with space. 



The distinction here indicated was long ago pointed 

 out and emphasized by Kant in his " Metaphysical 

 Foundations of Natural Science," where he speaks repeat- 



* "Common Sense of the Exact Sciences " (N. Y. Ed.), P 50. 



